1: \begin{abstract}

2:  We address the issue of numerical convergence in cosmological smoothed particle hydrodynamics simulations using 

3:  a suite of runs drawn from the \eagle{} project. Our simulations adopt subgrid models that produce realistic galaxy

4:  populations at a fiducial mass and force resolution, but systematically vary the latter in order to study their

5:  impact on galaxy properties. We provide several analytic criteria that help guide the selection of gravitational

6:  softening for hydrodynamical simulations, and present results from runs that both adhere to and deviate from them.

7:  Unlike dark matter-only simulations, hydrodynamical simulations exhibit a strong sensitivity to gravitational

8:  softening, and care must be taken when selecting numerical parameters. Our results--which focus mainly on star

9:  formation histories, galaxy stellar mass functions and sizes--illuminate three main considerations. First, softening

10:  imposes a minimum resolved escape speed, $v_\epsilon$, due to the binding energy between gas particles. Runs that

11:  adopt such small softening lengths that $v_\epsilon \simgt 10\,{\rm km s^{-1}}$ (the sound speed in ionised

12:  $\sim 10^4\,{\rm K}$ gas) suffer from reduced effects of photo-heating. Second, feedback from stars or active

13:  galactic nuclei may suffer from numerical over-cooling if the gravitational softening length is chosen below a

14:  critical value, $\epsilon_{\rm eFB}$. Third, we note that small softening lengths exacerbate the segregation of

15:  stars and dark matter particles in halo centres, often leading to the counter-intuitive result that galaxy sizes

16:  {\em increase} as softening is reduced. The structure of dark matter haloes in hydrodynamical runs respond to

17:  softening in a way that reflects the sensitivity of their galaxy populations to numerical parameters.

18: \end{abstract}

19: